sfxLbnlExt icon


Demand Forecasting with LBNL-4944E
sfxLbnlExtAxon funcs

Registered StackHub users may elect to receive email notifications whenever a new package version is released.

There are 5 watchers.


lbnl4944e(data, state: {})

Build the LBNL-4944E data model. This function takes a data grid containing the load, temperature and occupancy historical data, and returns a data model similar to the one in LBNL-4944E. The most notable difference is that occupancy is modeled using a 0 or 1 instead of creating a separate model for both modes.

The state parameter is a dictionary which may contain the following keys for configuring LBNL:

  • intervals: The number of temperature intervals to use. (default: 6)
  • minTemp: The minimum temperature to use for computing temperature

    intervals. By default, we find the min temperature of the temp column.

  • maxTemp: The maximum temperature to use for computing temperature intervals. By default we find the max temperature of the temp column.
lbnlDemo(site, train: pastMonth(), forecast: thisWeek())
lbnlForecast(model, dates: thisWeek(), opts: {})

Perform LBNL-4944E forecasting using a model trained by lbnlTrain for the given dates. Returns a grid containing the recent power, forecasted power, and occupancy. The opts parameter currently supports the following options:

  • backcast: set this marker tag to do "backcasting"
lbnlTrain(site, dates: pastMonth())

Get a trained model using the methodology outlined in LBNL-4944E.

  • site: The site to train the model against. Must have a site meter

    with a power sensor point.

  • dates: Use historical data from this date range for training the model.

The model metadata will include the computed cvrsme and nmbe metrics.

Published by SkyFoundry

Products & Services by SkyFoundry

Packages by SkyFoundry

Free packages